Development of toxic leadership scale in hospitality firms andanalysis of organizational citizenship behavior in the contextof toxic leadership, leader-member exchange, job stress,emotional exhaustion variables by data mining method

The aim of this study is to develop a valid and reliable toxic leadership scale for hospitality firms. This study will also determine the effect/effects of toxic leadership perceptions, leader member exchange, job stress, and emotional exhaustion levels in predicting the organizational citizenship behavior of hospitality employees. The research was carried out on two sample groups of employees working in four-and five-star hotels in Alanya. Both groups have similar characteristics. There were 336 participants in the first sample group and 623 participants in the second sample group. In order to test the construct validity and reliability of the toxic leadership scale, an exploratory factor analysis, and 27% sub-upper group comparisons were performed on the first sample group's data. A confirmatory factor analysis, a similar scale validity analysis, and a differential item functioning analysis were performed on the second sample group's data. The reliability of the scale was tested means of Cronbach Alpha coefficients and the Spearman-Brown split-half correlation. The SPSS, Lisrel, and jMetrik programs were used in these tests. A decision tree analysis was performed on the Weka program to determine which variables have an impact on employees' organizational citizenship behavior. The external validity of these results was tested with an artificial neural network analysis carried out using the Mathlab program. As a result of the analysis of the psychometric properties of the toxic leadership scale, it was determined that the scale has high construct validity and reliability. The scale consists of 27 items of a five-component structure, including narcissism, abusive supervision, authoritarian leadership, self-promotion and unpredictability. Based on the results of the decision tree and artificial neural networks analyses, it was determined that the variables with the largest impact on employees' organizational citizenship behavior are leader-member exchange, emotional exhaustion, toxic leadership, and job stress, in that order. Employees who exchange well with their leaders have high levels of citizenship behavior. In contrast, employees who exchange poorly with their leaders, are emotionally exhausted, and perceive their leaders as highly toxic have low levels of citizenship behavior.